Using a candidate gene approach, we propose to identify regions/loci of the human genome that make a polygenic contribution to the inherited risk of cancer. Approximately 100 candidates (300 markers) will be evaluated; these will represent loci or regions established as tumor suppressors, or encoding cell cycle control elements, DNA repair proteins, enzymes metabolizing xenobiotics, transcription factors implicated in oncogenic chromosome translocations, and regulators of apoptosis. This methodology is expected to be more efficient than random genome scanning because so many of the candidates may plausibly influence risk. We will collect sib pairs affected by any one of four important tumors demonstrating a previously described association with the HRAS1 VNTR--breast, colon, lung, and prostate cancer. To accept or reject candidates at the test loci, we will use the quantitative relationship in affected sibs between the extent of allele sharing and the relative risk (lambda s) of cancer. Our study design will permit the detection of such loci with nearly 80% power for lambda s is greater than or equal to 4 when the population prevalence of disease alleles is greater than or equal to 2% Because of genotyping costs we will analyze only one tumor at this level of detail. However, we will conduct an additional analysis with all four tumor types on the interaction of candidate loci with the HRASI VNTR. This component of the study design will permit the detection of genetic risk associated with positively interacting loci with nearly 100% power for lambda s is greater than or equal to 4; even for candidates with lambda s is 2, we will possess 70% power. Thus, we gain considerable power of detection, on the one hand, and the opportunity to observe new genetic interactions important to the genesis of common, "sporadic/nonfamilial" tumors, on the other. The large number of sib pairs required for this analysis will be obtained through a seven member consortium of institutions belonging to the Eastern Cooperative Oncology Group. As a consequence of these activities, and regardless of the outcome of our genetic studies, we will have collected a very large number of sib-pair samples, together with a broad-based, highly-organized, cancer genetic and clinical information resource that will serve as the basis for the establishment of a interactive genetic database for cancer-affected sibs.